Large-Scale Autonomous Gas Monitoring for Volcanic Environments: A Legged Robot on Mount Etna

TL;DR

Hazardous near-surface volcanic gas sampling motivates autonomous sensing. The authors demonstrate a legged ANYmal platform carrying a high-precision quadrupole mass spectrometer, integrated with a modular autonomy stack for long-range, terrain-aware operation on Mount Etna. Field tests include three autonomous missions achieving autonomy rates above $AR=90\%$ and several autonomous gas-source detections, plus a teleoperated plume-measurement mission that corroborates onboard measurements with handheld references. The study highlights practical lessons on adaptive sensing, tighter global-local planning integration, and hardware design needed to advance autonomous volcanology and source-proximal gas sampling in active volcanic environments.

Abstract

Volcanic gas emissions are key precursors of eruptive activity. Yet, obtaining accurate near-surface measurements remains hazardous and logistically challenging, motivating the need for autonomous solutions. Limited mobility in rough volcanic terrain has prevented wheeled systems from performing reliable in situ gas measurements, reducing their usefulness as sensing platforms. We present a legged robotic system for autonomous volcanic gas analysis, utilizing the quadruped ANYmal, equipped with a quadrupole mass spectrometer system. Our modular autonomy stack integrates a mission planning interface, global planner, localization framework, and terrain-aware local navigation. We evaluated the system on Mount Etna across three autonomous missions in varied terrain, achieving successful gas-source detections with autonomy rates of 93-100%. In addition, we conducted a teleoperated mission in which the robot measured natural fumaroles, detecting sulfur dioxide and carbon dioxide. We discuss lessons learned from the gas-analysis and autonomy perspectives, emphasizing the need for adaptive sensing strategies, tighter integration of global and local planning, and improved hardware design.

Figures (7)

• Figure 1: Aerial overview of Mount Etna with the four completed missions indicated. (A-B) Zoomed views showing detailed trajectories of two mission areas. (a-c) Images from autonomous helium-mapping missions. (d) Image from the teleoperated mission measuring volcanic gases at an active crater.
• Figure 2: Overview of the robotic system equipped with the gas-sensing payload and perception sensors. (a) Robot with labeled sensor components. (b) Gas-sensing payload with rollcage. (c) Transpector® MPH quadrupole mass spectrometer subsystem.
• Figure 3: Overview of software architecture. The architecture is organized into two main parts: the blue navigation modules, responsible for global and local path planning and locomotion control, and the green perception modules, which handle state estimation and elevation mapping. Information flows from high-level mission planning (top left) to low-level actuation (bottom left), while feedback from perception enables autonomous operation and live supervision of mapping and gas measurements. The numbers in pink correspond to the sections of this paper in which the respective modules are described.
• Figure 4: Computation of curvature-based adaptive lookahead. The robot’s position ($\mathbf{x}_{robot}$) is projected onto the path to obtain the nearest anchor point ($\mathbf{x}_i$), from which forward and backward secant vectors ($\mathbf{v}^{(p)}_i$, $\mathbf{v}^{(n)}_i$) define the local curvature ($\kappa_i$). The lookahead waypoint ($\mathbf{x}_{\mathrm{goal}}$) is then placed based on curvature-dependent lookahead distance ($L(\kappa_i)$).
• Figure 5: Results of the three autonomous field missions on Mount Etna: M1-Crater Rim, M2-Crater Descent, and M3-Volcanic Desert. For each mission: (a) Satellite image showing the mission area with marked interventions and detected gas source locations. (b) Terrain mesh with the traversed path, color-coded by relative gas concentration. (c) Time series of gas measurements using the same color map. (d-g) Representative images illustrating the robot’s operation in varying terrain conditions.